Large and complex IT service management environments that provide multiple services to a plurality of customers can create an unmanageable number of entities. It is best practice to make use of a logically centralized repository for the storage and access of the data, commonly referred to as a CMDB. The configuration data stored in this CMDB includes a representation of managed resources; such a representation is called a CI. CMDB CIs represent an extensive range of logical and physical entities and their complex relationships. A typical implementation of a CMDB may contain more than 800 abstract object classes, with implicit and explicit relationships between them that may be extended in an open-ended fashion.
Organizing entities into dependency trees or graphs for a high-level view of the topology eases systems management. A CMDB is broad and semantically rich enough that it may apply to higher layers such as, for example, a business process or a distributed application. A CMDB is also granular enough to represent, for example, tables in a database or enterprise Java beans (EJBs) used in an enterprise application. In real-world enterprise systems, there may be tens of thousands or more entities with complex relationships between them. Compositions are ideally suited to multi-layered topologies.
The CMDB records the existence, attributes, relationships, history and status of CIs. An attribute is a descriptive characteristic of a CI such as, for example, make, model, serial number, or location. A relationship describes associations, such as, for example, the dependency and/or connectivity between CIs.
At present there is little control over the manner in which these CIs are controlled, in particular the circumstances under which they can be changed as reflected by the lifecycle state that the CI is in. Because of this, it is possible for the administrators of this managed IT environment to make changes to configuration items in the CMDB at times that such changes should not be made, resulting in CI representations that are not correct in accordance with the administrators expectations. This leads to incorrect understanding by the administrators as to the correct state of the CMDB.